Hi,
Kian here,

I am new to this forum I hope this is the correct topic. Anyways, I am a ME with little exposure to the electronic & PCB design arena. That said, I am self taught and attempting to create a magnetic sensor board that will have a header with power input, GND and a signal output.

I am building this board around Honeywell's HMC1501 with hopes of sensing rotary motion. During my undergrad I took Circuits I -- and committed op-amps to my short term memory to scoot through exams and homework. Now it is back to haunt me...

I will provide the diagram provided by HMC1501 and my proposed circuit design built with EasyEDA.




The Op-Amp I chose was Gainsil GS8333-TR, and I *think* I can get away with powering it via single supply. The op-amp has 5-pins, namely, +in,-in,VSS,VDD and out. Where my confusion begins is VSS and VDD and how/where to assign those pins. From my experience VSS and VDD stem from the MOSFET family.

I will keep it short and sweet so I don't drag on with this. Any missing details I will be happy to provide just ask!

I am looking for feedback or perhaps guidance to getting my first ever PCB made!

 

Where my confusion begins is VSS and VDD and how/where to assign those pins.

Pin 5 of the op amp should connect to 5 volts.
R3 needs to connect to a 2.5 volt reference provided by the added voltage divider of R5 and R6.

 

Pin 5 of the op amp should connect to 5 volts.
R3 needs to connect to a 2.5 volt reference provided by the added voltage divider of R5 and R6.

I appreciate the reply! I will add those corrections.

EDIT:

I read in the spec sheet for the op-amp that VDD (Pin 5) should be connected to a 100nF ceramic capacitor... hence the free floating capacitor. I wasn't sure if that is typical or necessary to implement.

 

Yes, add the capacitor

 

The 100nF capacitor needs to be physically close to the op-amp, from pin 5 to 2. Rather than add R5 and R6 consider using 499K for R3 to ground (as in your first schematic), and add a 499K from +5V to op-amp pin 3. This better balances the amplifier than the R5-R6 solution. Also, to keep the filtering in balance add a 10nF cap from op-amp pin 3 to gound.

 

This better balances the amplifier than the R5-R6 solution.

That's not what the spec sheet shows.

 

@RPLaJeunesse is right and the spec sheet isn't.

Vdd and Vss are silly terms. They seem to have appeared along with 4000 CMOS, but in 4000 CMOS the power supply pins are both connected to the source of a MOSFET and neither is connected to the drain.
Could they have started with NMOS logic? In which case, how were PMOS logic power supplies labelled?

 

The spec sheet shows 249K to +2.5V. Consider that using two 10K resistors adds 5K to the 249K used (study Thevenin equivalents!), so you no longer have 249K to 2.5V but rather 254K. Using two 499K resistors, one to ground and one to +5V, gives 2.5V with a 249.5K equivalent resistance. A much better match to R4, and a better balance.

 

A much better match to R4, and a better balance.

I doubt if it really matters, R3 can return to ground seeing that the outputs from the sensor are at 1/2 supply.

 

Consider that using two 10K resistors adds 5K to the 249K used (study Thevenin equivalents!), so you no longer have 249K to 2.5V but rather 254K.

I see... essentially with the new solution I am putting the resistors directly in series

Could they have started with NMOS logic? In which case, how were PMOS logic power supplies labelled?

In the original description I had mentioned a MOSFET and was wondering how the VDD and VSS will play a part in the op-amp's power. Currently, I do not have a MOSFET tied into the circuit. I am just now returning to work on this project a bit more and read specs more carefully.

@RPLaJeunesse is right and the spec sheet isn't.

That said, I should update my circuit schematic with the proposed solution and added capacitors.

 

I doubt if it really matters, R3 can return to ground seeing that the outputs from the sensor are at 1/2 supply.

It matters if you want the sensor to respond to bipolar excitation (+/-45 degrees max, +/-60mV from sensor). It may matter if you want the sensor to respond at zero excitation (0 to 45 degrees max). And using 5V to bias the op-amp output at +2.5V only makes sense if the same 5V supply is the reference for the ADC and power for the sensor. If the ADC reference is other than +5V the use that to feed the upper 499K resistor, and to power the sensor as well. Keep it all ratiometric.

 

It matters if you want the sensor to respond to bipolar excitation (+/-45 degrees max, +/-60mV from sensor).

That's true you still need the 2.5 volt reference. If you really want to get technical use a LM336-2.5

 

That's true you still need the 2.5 volt reference. If you really want to get technical use a LM336-2.5

Actually, you don't want the bias point to be 2.5V. It should be at half the ADC full scale. If that is based on the 5V supply then the bias point should be 1/2 the 5V supply which, since the 5V supply can be other than 5.000V, means the bias point should not be held at 2.500V by an additional reference. Keeping everything ratioed to just one "fixed" voltage works best!

 

Thanks you guys. You have collectively provided me with some homework. I am going to review what you guys have suggested and come back with any questions.

If you really want to get technical use a LM336-2.5

I am actually not confident in the op-amp I selected. I just looked for one that seemed reasonable. I will check out the specs for the LM336-2.5. I just know I am looking to amplify my signal coming off the bridge and make it radiometric from some minimum voltage to some maximum voltage. Such that my controller can convert the signal into a angular-locational human-readable value.

 

I am actually not confident in the op-amp I selected.

I don't see anything wrong with that op amp, however an instrumentation chip such as an INA122 is a better choice then the GS8333-TR .
The LM336-2.5 is a voltage reference chip, what is the reference voltage on the ADC going to be?

 

I don't see anything wrong with that op amp, however an instrumentation chip such as an INA122 is a better choice then the GS8333-TR .
The LM336-2.5 is a voltage reference chip, what is the reference voltage on the ADC going to be?

Reviewing the specs of my mainboard, I will need to design around 3.3v power input. Reference voltage referring to minimum voltage output from the sensor? I would imagine 2.5v would make sense, though, I am not totally sure at this point.

The closest thing to this project I have gotten was working with a pressure transducer, and converting it's data via Arduino IDE and plotting it through MATLAB. It was an analog ratiometric sensor, however, this project is quite a step up from that seeing as though I need to make a daughterboard for my mainboard.

More descriptively, I am looking to make a magnetic filament monitor for a 3D printer. The tech that worked for my mainboard before is out of stock and hasn't been in stock since 2018 so I have tasked myself with coming up with my very own for my personal use. I plan on buying the components once I am confident to do so and use my labs oscilloscope to test its output before doing a PCB.

 

Reference voltage referring to minimum voltage output from the sensor?

No. The ADC input on the micro uses a reference voltage, it's normally the same as the supply voltage which you mentioned at 3.3 volts. As mentioned in previous post the reference voltage on the amplifier chip should be at 1/2 the supply voltage.
Using the INA122:

 

Note that the INA122 REF input needs a VERY low-Z drive. So add a pair of resistors off +3.3V supply and buffer with an op-amp to drive the REF pin. INA122 just complicates things, doesn't really gain you anything. Go back to the earlier design with dual 499Ks, look at using the MCP6021 op-amp, and keep everything ratiometric to the (now 3.3V) supply.

 

INA122 just complicates things, doesn't really gain you anything.

Actually the INA122 is made for this type of application. Use a Texas instruments REF2033, for the 1.65 volt reference offset.
One chip, one reference IC and one resistor, I don't see how that complicates things.
I'm sure RPLaJeunesse will disagree but this is my recommendation. The TS can decide for himself.